1. Predictive Analysis

   1. predict income
   2. Max achieved so far: 45.6% accuracy

   Try to increase accuracy - better oversampling, folding, validation, tuning, cross validation.

2. Comparative Analysis

   1. private vs anonymised data
      1. Logistic Regression
      2. SVC
      3. Random Forest/ Descision Tree

3. Coefficient Optimisation with Genetic Algorithm

   1. PyGAD
   2. Temporal Attack with AI
      1. generate data releases
      2. attack private table:
      3. attack anonyised table:

TEMMPORAL ATTACK ON ANONYMISED TABLES

1. Customised Activation Functions

   ReLu——> customised——→ f(correlation)——>

we adjsut the rate of growth of our mReLu curve using the values obtained after getting correlation between each pair of attributes.

1. get correlation

2. model math function for the lambda

3. Try to recursively link records with appended knowledge of previous records.